|Document Type:||Contract Report|
|Title:||Genetic stock identification|
|Author/Editor:||George B. Milner, David J. Teel, P. B. Aebersold, Fred M. Utter|
|Publisher:||National Marine Fisheries Service|
|Contracting Agency:||Bonneville Power Administration. Portland, Oregon|
The results of the first year's investigation of a 5-year plan to demonstrate and develop a coast-wide genetic stock identification (GSI) program are presented. The accomplishments under four specific objectives are outlined below:
1. Improved Efficiency through Direct Entry of Electrophoretic Data into the Computer. A program is described that was developed for direct computer entry of raw data. This program eliminated the need for key–to–tape processing previously required for estimating compositions of mixed fisheries, and thereby permits immediate use of collected data in estimating compositions of stock mixtures.
2. Expand and Strengthen Oregon Coastal and British Columbia Baseline Data Set. Electrophoretic screening of approximately 105 loci of samples from 22 stocks resulted in complete data sets for 35 polymorphic and 19 monomorphic loci. These new data are part of the baseline information currently used in estimating mixed stock compositions.
3. Conduct a Pilot GSI Study of Mixed Stock Canadian Troll Fisheries off the West Coast of Vancouver Island. A predominance of lower Columbia River (fall run), Canadian, and Puget Sound stocks was observed for both 1984 and 1985 fisheries. Stocks other than Columbia River, Canadian, and Puget Sound contributed an estimated 13 and 5%, respectively, to the 1984 and 1985 fisheries.
4. Valid ation of GSI for Estimating Mixed Fishery Stock Composition. Baseline data from the Columbia River southward were used to simulate northern and central California fisheries. These simulations provided estimates of accuracy and precision for mixed sample sizes ranging from 250 to 1,000 individuals. Sacramento River stocks had a heavier weighting in the central (89%) than in the northern (25%) fishery. Accuracy and precision increased for both fisheries as sample sizes increased and also were better for those estimates that were over 5%. Extrapolations from these estimates indicated that sample sizes of 2,320 and 2,869 would be required to fulfill coefficients of variation (SD/estimated contribution) of 20% with respective confidence intervals of 80 and 95% in stock groupings of the northern fishery. Similarly, sample sizes of 2,450 and 3,030 would be required in the central fishery.
A concluding section noted that these investigations are part of an effort involving many agencies. The requirements for simulation preceding actual sampling of stock mixtures and for continued monitoring and development of baseline data sets were emphasized.